Sociodemographic differences in diabetic retinopathy screening; using patient-level primary care data for health equity audit

Simon DS Fraser¹, GE Watkinson², CA Rennie³, D King², H Sanderson², L Edwards⁴, P Roderick^1
1Public Health Sciences and Medical Statistics, University of Southampton, Southampton General Hospital, Southampton, ²NHS Southampton City, Southampton, ³Southampton Eye Unit, Southampton General Hospital, Southampton, ⁴NHS Hampshire, Eastleigh, Hampshire, UK

Background: The prevalence of diabetes is increasing worldwide and there is inequality in the distribution of diabetic complications. Diabetic retinopathy is the leading cause of blindness in adults of working age in the UK, and certain risk factors are recognized. Retinopathy screening in the UK involves annual digital retinal photography and image grading. Auditing equity in retinopathy screening poses unique challenges, and screening program data are often incomplete for variables relevant to equity. Using two sources of patient-level primary care data, we conducted a health equity audit comparing the access and uptake of screening between groups of people with diabetes in each of three screening programs covering this area of southern England.
Methods: A patient-level dataset using data from general practices and a combined health record was used to compare dimensions of equity (gender, age, length of time since diabetes diagnosis, type of diabetes, presence of hypertension, socioeconomic deprivation, ethnicity, and screening program) between people with and without a record of retinopathy screening within three years in Hampshire and the Isle of Wight, UK.
Results: Anonymized data for 70,004 people with diabetes were obtained from 205 (88%) general practices. In total, 62,836 people (89.8%) had a record of screening within three years and 7168 (10.2%) did not. Lower uptake of screening was independently associated with the youngest and oldest age groups (compared with 50–79-year-olds), recent diabetes diagnosis, and deprived areas. Diagnosed hypertension was positively associated with screening.
Conclusion: Evaluating equity in screening programs is important to help reduce inequalities. We found evidence of inequity in access and uptake of retinopathy screening. Primary care data contained more information than screening program data. Using a combined health record was more efficient than obtaining data directly from general practices, but data were incomplete for deprivation measures at the time of this audit. Our audit informed subsequent efforts to improve equity in local diabetic retinopathy screening services.

Keywords: inequality, diabetes, eye, complications, screening